My first attempt at a male Lora, was trained on SD 1.5, but I was using Hyper v2 to generate renders. This model was created using publicly available images.

I have found the LoRA weight works in the 1.0-1.2 range. Admittedly this Lora struggles sometimes especially straight on to get the resemblance right. Also, I may have had too few photos of Penn smiling, since serious faces are more accurate.

Disclaimer: This AI model, generated for the purposes of artistic expression and technological demonstration, is not intended nor was it designed to produce illicit or inappropriate imagery. As the creator, I strongly discourage and disapprove of the use of this work for creating or distributing potentially immoral, unethical, or illegal content. By choosing to download this model, you agree to these conditions and undertake the responsibility not to misuse this work in any way contrary to these intentions.